Isn't it a typo?

Isn't it a typo? In s2n/crypto/s2n_hash.c

138    case S2N_HASH_SHA384:
139        eq_check(size, SHA384_DIGEST_LENGTH);
140        r = SHA512_Final(out, &state->hash_ctx.sha384);
141        break;

at line 140

r = SHA384_Final(out, &state->hash_ctx.sha384);
       ^^^

In OpenSSL, it's same meaning.

int SHA384_Final(unsigned char *md, SHA512_CTX *c)
{
    return SHA512_Final(md, c);
}

[raycoll]

Currently this is functionally correct, but I think switching to SHA384_Final() makes more sense for readability.

It also makes sense for consistency: s2n uses SHA384_Update() in s2n_hash.c:

case S2N_HASH_SHA384:
        r = SHA384_Update(&state->hash_ctx.sha384, data, size);
        break;

and openssl does the same delegation delegation to the SHA512 function for the SHA384_Update() call:
https://github.com/openssl/openssl/blob/349807608f31b20af01a342d0072bb92e0b036e2/crypto/sha/sha512.c#L232